Most turnstile tenders still compare photographs, steel grades and price per lane. Three engineering decisions determine whether a lane runs for a decade or becomes a maintenance liability: the drive train, the sensor array that counts people, and the failure logic when power, network or credentials disappear.
A solenoid lock is cheap and fast to open, but it slams: every closure is a mechanical impact that drives warranty claims in high-traffic lanes. Brushless DC servo drives with an encoder close under position control: the controller knows the arm angle, decelerates before the end stop, and reverses instantly when a beam is broken. The measurable difference is MCBF (mean cycles between failures) and acoustic signature — a servo lane is quiet at 60 passages per minute, a solenoid lane is not.
| Gate type | Typical drive | MCBF (indicative) | Open time | IR beam pairs | Throughput |
|---|---|---|---|---|---|
| Tripod turnstile | Solenoid or stepper, drop-arm | 3–5 million cycles | 0.5–1.0 s | 3–6 | 20–30 /min |
| Swing barrier | Brushless DC servo + encoder | 5–8 million cycles | 0.4–0.6 s | 6–12 | 30–40 /min |
| Speed gate | Brushless servo, closed-loop | 8–10 million cycles | <0.4 s | 12–24 | 40–60 /min |
| Full-height gate | Servo or gear motor, dual wing | 5–8 million cycles | 0.8–1.2 s | 10–16 | 15–25 /min |
Adding beam pairs does not automatically stop tailgating. What matters is the tracking algorithm: how the controller associates a sequence of beam breaks into one person, how it distinguishes a wheelchair or a rolling suitcase from a second body, and how quickly it decides to close. A 24-beam array with naive logic false-alarms on luggage; a 12-beam array with per-cell occupancy tracking and a calibrated profile threshold does not. Ask for the false-rejection rate measured on a mixed test group — adults, children, wheelchair users, trolley bags — over at least 1,000 passages.
| Scenario | Correct behaviour | Common failure |
|---|---|---|
| Second person enters behind valid user | Alarm, hold wing open, log event with snapshot trigger | Wing closes on the follower — injury risk |
| Trolley bag trails the user | Recognised as same passage, no alarm | False tailgate alarm, lane stoppage |
| User reverses mid-lane | Re-open and re-validate, no lock-out | Wing traps user, manual release needed |
| Network down | Offline whitelist runs locally, events buffered | Lane fails closed, queue builds |
| Fire alarm active | All wings free-wheel, arms drop, no power required | Software-only release that needs mains |
Every lane needs a documented answer to four questions: what happens on power loss, on network loss, on fire signal, and on forced push. Fail-safe (open) suits life-safety egress; fail-lock suits asset protection. Neither is universally correct, so the lane must support both and be configurable per installation. Fire release must be hardware-level — a relay dropping power to the brake, not a command sent over TCP/IP that depends on a running controller.
| Project / date | Scale (public data) | What it demands from a lane |
|---|---|---|
| EU Entry/Exit System progressive start 12 Oct 2025; fully operational 10 Apr 2026 |
29 European countries; biometric registration of non-EU short-stay travellers | Anti-tailgating to border standard, fingerprint and face capture, documented MTBF, audit logs per crossing |
| Riyadh Metro opened 1 Dec 2024 |
6 lines, 176 km, 85 stations | Fare gates holding peak surges in high ambient temperature; dust ingress protection; remote diagnostics across 85 stations |
| Expo 2025 Osaka 13 Apr – 13 Oct 2025 |
184 days of continuous operation | Zero-tolerance for lane downtime; fast credential mix (QR, RFID, face); multilingual display and staff override |
| FIFA World Cup 2026 11 Jun – 19 Jul 2026 |
16 stadiums, 48 teams, 104 matches | 25–40 validations per minute per lane before kick-off; offline fail-open; real-time counts to command centre |
Most turnstile tenders still compare photographs, steel grades and price per lane. Three engineering decisions determine whether a lane runs for a decade or becomes a maintenance liability: the drive train, the sensor array that counts people, and the failure logic when power, network or credentials disappear.
A solenoid lock is cheap and fast to open, but it slams: every closure is a mechanical impact that drives warranty claims in high-traffic lanes. Brushless DC servo drives with an encoder close under position control: the controller knows the arm angle, decelerates before the end stop, and reverses instantly when a beam is broken. The measurable difference is MCBF (mean cycles between failures) and acoustic signature — a servo lane is quiet at 60 passages per minute, a solenoid lane is not.
| Gate type | Typical drive | MCBF (indicative) | Open time | IR beam pairs | Throughput |
|---|---|---|---|---|---|
| Tripod turnstile | Solenoid or stepper, drop-arm | 3–5 million cycles | 0.5–1.0 s | 3–6 | 20–30 /min |
| Swing barrier | Brushless DC servo + encoder | 5–8 million cycles | 0.4–0.6 s | 6–12 | 30–40 /min |
| Speed gate | Brushless servo, closed-loop | 8–10 million cycles | <0.4 s | 12–24 | 40–60 /min |
| Full-height gate | Servo or gear motor, dual wing | 5–8 million cycles | 0.8–1.2 s | 10–16 | 15–25 /min |
Adding beam pairs does not automatically stop tailgating. What matters is the tracking algorithm: how the controller associates a sequence of beam breaks into one person, how it distinguishes a wheelchair or a rolling suitcase from a second body, and how quickly it decides to close. A 24-beam array with naive logic false-alarms on luggage; a 12-beam array with per-cell occupancy tracking and a calibrated profile threshold does not. Ask for the false-rejection rate measured on a mixed test group — adults, children, wheelchair users, trolley bags — over at least 1,000 passages.
| Scenario | Correct behaviour | Common failure |
|---|---|---|
| Second person enters behind valid user | Alarm, hold wing open, log event with snapshot trigger | Wing closes on the follower — injury risk |
| Trolley bag trails the user | Recognised as same passage, no alarm | False tailgate alarm, lane stoppage |
| User reverses mid-lane | Re-open and re-validate, no lock-out | Wing traps user, manual release needed |
| Network down | Offline whitelist runs locally, events buffered | Lane fails closed, queue builds |
| Fire alarm active | All wings free-wheel, arms drop, no power required | Software-only release that needs mains |
Every lane needs a documented answer to four questions: what happens on power loss, on network loss, on fire signal, and on forced push. Fail-safe (open) suits life-safety egress; fail-lock suits asset protection. Neither is universally correct, so the lane must support both and be configurable per installation. Fire release must be hardware-level — a relay dropping power to the brake, not a command sent over TCP/IP that depends on a running controller.
| Project / date | Scale (public data) | What it demands from a lane |
|---|---|---|
| EU Entry/Exit System progressive start 12 Oct 2025; fully operational 10 Apr 2026 |
29 European countries; biometric registration of non-EU short-stay travellers | Anti-tailgating to border standard, fingerprint and face capture, documented MTBF, audit logs per crossing |
| Riyadh Metro opened 1 Dec 2024 |
6 lines, 176 km, 85 stations | Fare gates holding peak surges in high ambient temperature; dust ingress protection; remote diagnostics across 85 stations |
| Expo 2025 Osaka 13 Apr – 13 Oct 2025 |
184 days of continuous operation | Zero-tolerance for lane downtime; fast credential mix (QR, RFID, face); multilingual display and staff override |
| FIFA World Cup 2026 11 Jun – 19 Jul 2026 |
16 stadiums, 48 teams, 104 matches | 25–40 validations per minute per lane before kick-off; offline fail-open; real-time counts to command centre |